Ought and FHI's AI Safety Group are collecting data on how people come to judgments over time.

Start playing here (no sign-up required)!

Take part and play games about:

• Fermi arithmetic problems, or
• Fact-checking political statements (decide if a statement is “fake news”), or
• Deciding how much you like an academic paper in Machine Learning

You’ll get feedback on your progress over time and your calibration.

Want to know more about the motivation behind collecting this data, and why we expect it will be useful for progress on AI Safety? Check out our thoughts here.